Hello,
Iám Rainer and Iám new in the diy Audio Topic.
I make altimes projects with a group of young people.
The last time very much with Arduino in all directions.
We also have made a lot of PCBs and order them form JLPCB. Perfect.
Now the young pupils and I want to do a own Audio Amp.
We read a lot of TDA7293 and also LM3886
We decide of LM3886 and pic out some schematics and try to combine this to
a of our option usefull own schematic.
For me is very important that we got a first good and working result for the young pupils.
So i go the way to this forum and looking for help.
..... Nothing is more bad as unlucky eyes of childs if that what we do not work.....
So I want to ask kindly if I get here the help and the members here have a look of my schematic ,which I want to share and looks for erros?
Iám Rainer and Iám new in the diy Audio Topic.
I make altimes projects with a group of young people.
The last time very much with Arduino in all directions.
We also have made a lot of PCBs and order them form JLPCB. Perfect.
Now the young pupils and I want to do a own Audio Amp.
We read a lot of TDA7293 and also LM3886
We decide of LM3886 and pic out some schematics and try to combine this to
a of our option usefull own schematic.
For me is very important that we got a first good and working result for the young pupils.
So i go the way to this forum and looking for help.
..... Nothing is more bad as unlucky eyes of childs if that what we do not work.....
So I want to ask kindly if I get here the help and the members here have a look of my schematic ,which I want to share and looks for erros?
Post your schematic, but the most reliable source for a circuit would be
the manufacturer's own data sheet. See figure 3, on page 6.
Don't forget the heat sink! I would suggest using the isolated package, if available.
https://www.ti.com/lit/ds/symlink/l...https%3A%2F%2Fwww.ti.com%2Flit%2Fgpn%2Flm3886
the manufacturer's own data sheet. See figure 3, on page 6.
Don't forget the heat sink! I would suggest using the isolated package, if available.
https://www.ti.com/lit/ds/symlink/l...https%3A%2F%2Fwww.ti.com%2Flit%2Fgpn%2Flm3886
Last edited:
Be sure you can get supply of known-good chips. Recent global disruption has made many chips out-of-stock. Scammers have been selling chips that don't work right.
Note that as a student project, a 1 Watt or 3 Watt chip may be as much fun as the 50+W monsters, and are not in such demand. Car-sound 12V chips are ubiquitous and will make 15W in 4 Ohms, at very safe voltages.
Last edited:
Hello,
the decision is fixed. We have already made our first own designs and our own
boards with the LM3886, which work great.
We have a source where we got all the original components safely.
We are actively supported here, and there is also an EMC laboratory nearby which supports us and and the FAE there can measure the THD for us
with a spectrum analyzer.
Then we also have support from a HW developer (digital technology), but he is not an expert in Amps and analog technology.
But in generall he support us, so much he can.
------> Now we want to improve the THD and linearity of the LM3886.
I would now like to share our design draft with you.
I ask for assessments whether this principle of the composite mode can work?
Please don't stick to the component sizes and values just yet, it's all about the hardware principle of the circuit.
So thank you in advance and I look forward to all constructive help.
best regards
Rainer
the decision is fixed. We have already made our first own designs and our own
boards with the LM3886, which work great.
We have a source where we got all the original components safely.
We are actively supported here, and there is also an EMC laboratory nearby which supports us and and the FAE there can measure the THD for us
with a spectrum analyzer.
Then we also have support from a HW developer (digital technology), but he is not an expert in Amps and analog technology.
But in generall he support us, so much he can.
------> Now we want to improve the THD and linearity of the LM3886.
I would now like to share our design draft with you.
I ask for assessments whether this principle of the composite mode can work?
Please don't stick to the component sizes and values just yet, it's all about the hardware principle of the circuit.
So thank you in advance and I look forward to all constructive help.
best regards
Rainer
Your circuit consists of an input buffer, which might be ok. Second, it aims at correcting the output of the 3886 with an op-amp. Simple said, you add a second active over all feedback to it.
Did you ever listen to a simple, correct build of this chip amp? I fear you didn't. If so, you would not think over improving THIS amp, but try to build better loudspeakers.
Even with very good speakers, I do not like to write "high end", this amp is one of the best A/B and chip amps ever made. It has a surprising tolerance to any stupid situation, like shortcuts under full load or supply voltages failing and does not plop during on/ off switching. Just build right.
If you talk about "amp sound" you usually talk about faults in reproduction, caused by the amp. A perfect amp does not sound, so you can not point to some audible effect and say "this was the amp". The 3886 comes as close to a perfect amp as one can whish. At, maybe to little cost.
Most "bad" amps start to sound stressed if you turn up the volume, the 3886 simply gets louder. On small, very good speakers, you can even experience how the chassis start to compress, but without distortion caused by the amp.
It will be hard to find a loudspeaker that is able to reproduce any improvement over the 3886's reverence design.
Even the measurements are fantastic.
There is Tom Christensen, who did extensive work around this chip. Search for "3886 done right". Maybe you can find inspiration in his designs, but please realize: what he did was not the result of an afternoon thinkering, but a lot of full time work.
If you are not a very gifted developer with measuring gear worth a fortune, your additional feedback loop may not improve anything. I fear it will degrade the perfect sound and only make the amp unstable. You are overengineering it and are trying to fix something not broken. If you design has problems with the THD and linearity, I'm pretty sure your "improvements" or layout cause them. Again, read what Tom wrote.
Sorry I can not help any more and as you already made the PCB, there isn't much to help anyhow.
My advice: Keep the buffer if really needed and stick to the reverence design with the rest.
There is even a very cheap (15€) version on Amazon/ eBay which already shows you how good this amp can be. Done about right. May not be the best it can get, but working and reliable. Add a decent power supply and compare it to your design. You may be surprised.
Did you ever listen to a simple, correct build of this chip amp? I fear you didn't. If so, you would not think over improving THIS amp, but try to build better loudspeakers.
Even with very good speakers, I do not like to write "high end", this amp is one of the best A/B and chip amps ever made. It has a surprising tolerance to any stupid situation, like shortcuts under full load or supply voltages failing and does not plop during on/ off switching. Just build right.
If you talk about "amp sound" you usually talk about faults in reproduction, caused by the amp. A perfect amp does not sound, so you can not point to some audible effect and say "this was the amp". The 3886 comes as close to a perfect amp as one can whish. At, maybe to little cost.
Most "bad" amps start to sound stressed if you turn up the volume, the 3886 simply gets louder. On small, very good speakers, you can even experience how the chassis start to compress, but without distortion caused by the amp.
It will be hard to find a loudspeaker that is able to reproduce any improvement over the 3886's reverence design.
Even the measurements are fantastic.
There is Tom Christensen, who did extensive work around this chip. Search for "3886 done right". Maybe you can find inspiration in his designs, but please realize: what he did was not the result of an afternoon thinkering, but a lot of full time work.
If you are not a very gifted developer with measuring gear worth a fortune, your additional feedback loop may not improve anything. I fear it will degrade the perfect sound and only make the amp unstable. You are overengineering it and are trying to fix something not broken. If you design has problems with the THD and linearity, I'm pretty sure your "improvements" or layout cause them. Again, read what Tom wrote.
Sorry I can not help any more and as you already made the PCB, there isn't much to help anyhow.
My advice: Keep the buffer if really needed and stick to the reverence design with the rest.
There is even a very cheap (15€) version on Amazon/ eBay which already shows you how good this amp can be. Done about right. May not be the best it can get, but working and reliable. Add a decent power supply and compare it to your design. You may be surprised.
Hello Rainer,
i did some work for about half a year with a composite design using the TDA7294 and have learned some by simulating an testing the results on the real circuit, so some comments from me about the circuit you posted:
I started a homepage for my design, but is not completed yet - if you are interested in my way of building a composite:
https://bebo65.webador.de
Bernd
i did some work for about half a year with a composite design using the TDA7294 and have learned some by simulating an testing the results on the real circuit, so some comments from me about the circuit you posted:
- R17 makes no sense to me and can be replaced by a piece of wire
- i assume C21 makes more trouble than it solves - it is not neccessary in my opinion on this place, a better place for it will be between Pin 9 and Pin 10 of IC2
- if your circuit is oscillating a higher value of C17 should solve the problem
- IC3 and the surrounding circuitry is not necessary, C19 does the same job
- use X7R Parts only for Power supply
- separate high current ground from low current ground
- the best point to connect the two grounds is the speaker output connection of your board
I started a homepage for my design, but is not completed yet - if you are interested in my way of building a composite:
https://bebo65.webador.de
Bernd
If this is important i suggest you don't chose a composite amp to start with, unless you want to build and test it before the young pupils start the project.
As turbowatch2 suggested have a look at tom christensens neurochrome website. I suppose you first should get the pcb and the grounding right before going for the composite adventure with kids.
Last edited:
i would have a look at a project called miniref thats posted on this forum. i successfully integrated 3886 into that circuit and it works and sounds marvelous